Having dealt with the question of why organisations need to develop agile global networks, we can now focus on the much more challenging question of how to develop, deploy and manage agile global networks.
To do so, we have to start by looking at where most organisations are starting their journey towards agility (for a deeper dive on this subject, read TNC’s recent article HERE). For most multinationals, their agility journey started with a global network, most likely delivered using MPLS, and most likely delivered by a single global supplier. Back in these days, there weren’t many of these single global suppliers, so the supplier was probably one of five or six of the big names.
In those days, agility also wasn’t top of anyone’s must have list. The driving force was the concept that “we specialise in making widgets, so we should focus on making widgets”, and leave a telecoms company to deliver the network. That model ruled the waves for years and years, but the cracks started to show over recent years.
The main problems organisations started to find with their monolithic global networks was that they were clunky and expensive. Local country IT teams complained that they could buy an internet connection for half the price of the global WAN link, with twice the bandwidth and delivered in a third of the time. Traffic was increasingly heading to the internet but corporate WANs were built on MPLS. Service performance from the global telecoms companies was poor, and as demand for network services rose, and the need for agility increased, these clunky service models simply weren’t up to the job.
As a result, over the last 5 years, we’ve seen an explosion of new approaches. Some companies are throwing away their global WANs altogether and just buying lots of local internet connections. Others are moving to regional WANs from best-of-breed regional operators. Some others are using Systems Integrators to stitch together local services to something more globally unified. Whilst radically different, what all of these models are trying to do is find a balance between cost, performance and complexity.
Put simply, the more local the supplier, the cheaper they can offer connectivity in their specialised geography, and the quicker they can deploy. For example, a local ISP in Thailand would deliver an internet connection in Bangkok cheaper and quicker than a regional specialist, and the regional specialist would be cheaper and quicker than a global telecoms provider. However, a single global provider solution is easier to manage, and a network built using lots of local suppliers is much more complex.
The reality is that all of these models have some applicability, so the challenge for each organisation is to find the model that is most aligned to its requirements. In the following table, TNC shows the key characteristics of each model and the organisations to which each is suited.